PROTEUS: Evaluating the Use of Artificial Intelligence to Support Stress Echocardiography Testing for Heart Disease (PROTEUS)

PROTEUS: Evaluating the Use of Artificial Intelligence to Support Stress Echocardiography Testing for Heart Disease

PROTEUS: A PROspective Randomised Controlled Trial Evaluating the Use of Artificial Intelligence in Stress Echocardiology

University of Oxford, Academic Health Science Centres, National Institute for Health Research, United Kingdom

PROTEUS is a multicentre, two arm, randomised controlled trial of a medical device to assess the impact of the introduction of EchoGo into the standard care pathway for stress echocardiology.

Title: A Prospective Randomised Controlled Trial Evaluating the Use of AI in Stress Echocardiography Trial design: Multi-centre, two-arm, Randomised Controlled Trial of a medical device. Randomised 1:1 to either: Standard care (comparator), or Standard care plus AI Platform (EchoGo) Trial participants: Adults undergoing clinically indicated stress echocardiogram (SE) to assess inducible ischaemia at participating NHS Trusts. Having benchmarked the device performance (Upton 2019), and having successfully deployed and tested the required infrastructure, we now propose to prospectively test the software in clinical practice. Based on the existing data for accuracy we would expect use of the EchoGo Platform in the clinical pathway will have a significant benefit reducing inter-reader variability, increasing accuracy and thereby reducing health costs. To formally evaluate the impact of automated AI quantification and accuracy of the decision to make a referral to coronary angiogram, following stress echocardiography, we propose a prospective, multi-centre, randomised controlled trial that compares provision of an AI derived report to a clinician against clinician-only standard of care. This will provide the highest level of evidence of any potential benefit or disadvantages prior to adoption of the software for widespread clinical implementation.

A multi-centre, two-arm, randomised controlled trial. Randomised 1:1 Standard care group (comparator) Standard care plus AI platform (EchoGo)

Following the stress echocardiogram examination, participants will follow the care pathway determined by the clinical team managing their care. If the participant is assigned to the study intervention arm the EchoGo report will inform this decision making process. In order to minimise bias in any outcome assessment clinicians will be asked not to share the details of the EchoGo report received for participants within the intervention arm with the participant or with colleagues who may undertake future investigations eg: angiography.If a clinician subsequently considers there is a clinical need to know the details of the EchoGo report this can be made available at the discretion of the study team.

Patients will receive standard care plus their Echocardiogram will be sent to Ultromics for AI assessment. The report from the assessment will be sent to the clinician, and utilised to inform the patients further care.

EchoGo Pro is an software Medical Device which is able to assess Stress Echocardiograms and provide a binary classification of risk of CAD

Investigate if EchoGo plus standard care is non-inferior to standard care alone for aiding referral to coronary angiogram following stress echocardiogram (SE)

Outcome measure - AUROC for the ability to make an appropriate referral to coronary angiogram at 6 months following stress echocardiogram. Appropriate management will be defined as a composite of coronary angiography that demonstrates severe coronary disease fulfilling clinical care guidelines for revascularisation or decision for medical management/reassurance without unanticipated serious adverse cardiac events Outcome data will be collected from participant's medical records at 3 and 6 months.

Investigate if EchoGo plus standard care is superior to standard care alone in predicting the risk of severe cardiovascular disease following stress echocardiogram (SE)

Outcome shall be measured through - AUROC for the ability to make an appropriate referral to coronary angiogram.

Investigate the appropriate clinical management decisions made when using EchoGo plus standard care compared to standard care alone.

Outcome measure - Appropriate clinical management following an echocardiogram, for referring a patient to coronary angiogram, determined by adjudication committee review of follow up data.

Establish if clinical management decision making is affected by review of the EchoGo report as recorded by clinician self-report.

Clinician diagnostic confidence in their interpretation of the stress echo report will be measured by clinician self-report at baseline, 3 months and end of study at each participating site.

Outcome measured by Inter-clinician and in-site variability at baseline, 3 months and at end of study.

Establish if using Ultromics' EchoGo Platform affects the number of subsequent investigations for cardiovascular disease. for cardiovascular disease

Outcome shall be measured through collation of the number of incidences of the following (occurring up to 6 months following stress echo): Myocardial Perfusion Scan CT coronary angiogram Stress echo Invasive coronary angiogram Stress CMR Exercise tolerance test Initiation of anti-anginal medication/medical management of angina Data will be collected from participants medical records at 3 and 6 months

Establish if using Ultromics' EchoGo Platform affects patient reported coronary artery disease symptoms.

Outcome shall be measured through the collation of the number of coronary artery disease symptoms and impact on participant health status as measured by patient-reported short Seattle Angina Questionnaire (SAQ-7) at trial entry, 3 months and 6 months

Assess the health economics impact of implementation and use of Ultromics' EchoGo platform in NHS units

Inclusion Criteria: Willing and able to provide informed consent Male or female, ≥18 years of age at study entry Referred to an NHS Trust for stress echocardiography for investigation of ischaemic heart disease Exclusion Criteria: More than moderate valvular heart disease Left ventricular outflow tract obstruction defined as a gradient > 30mmHg (fixed or dynamic; supravalvular, valvular or sub-valvular) Significant co-morbidities (e.g. cancer) with an expected life-expectancy of under 12 months in the investigator's opinion Previous coronary artery bypass graft or other cardiac surgery Congenital or inherited myocardial disease